A deep learning model for bearing fault diagnosis based on convolution neural network with multi-channel and residual network

Traditional bearing fault diagnosis algorithms mostly rely on expert experience and prior knowledge, which can no longer meet the actual requirements of industrial big data. This paper proposes a new deep learning model that combines the multi-channel and wide first layer structures, and uses dropout technology, regularization, batch normalization, and other methods to solve the problem of overfitting in the network structure problem, and the introduction of the residual network to solve the problem of network degradation. Experimental results show that the model has an average accuracy of 100% in the bearing data set of Western Reserve University, showing good adaptive ability. The comparison results with mainstream diagnostic algorithms shows that the proposed method has good anti-noise abilit.